it makes @andabata with his 400 photos stacked look boring suddenly. and in the past I thought he was crazy!
2 Likes
@fab4space, It is a totally unexpected and counter-intuitive result. I understand the point about guiding (and tracking, I guess) and mount. Although I guess you’d get some image rotation over the course of the night, Siril would probably correct for that.
But noise. There are a number of sources of noise, Read noise, Shot noise and light pollution and they don’t all pull in the same direction. But if you are using an untracked mount, I’m guessing 1 or 2 secs is probably about right.
You’ve managed to get a fantastic amount of detail, and I imagine that’s your lovely 400 mm F/4.5 Dobsonian at work. I can’t get anything like that on my 200 mm F/5 Newtonian.
In my own images, I’ve gone up to 300 secs, and down to 30 secs. I find 180 s a happy medium, but tracked and guided. If I was unguided, I’d probably drop to 45 secs.
I’ve also experimented with ISO from 400 to 6400, and again, I find that 1600 is better.
My main battle at the moment is with fixed pattern noise (FPN) on my DSLR. I’m experimenting with dithering to remove that, but it’s a slow and painful process. Does your method help with FPN?
It’s always nice to have our preconceptions challenged and you’ve certainly done that.
Thanks and regards,
Steve.
Thanks @rbarbera @heckflosse @darix @Steve_Cookson for your comments 
@Steve_Cookson , with this altaz mount Siril handle the field rotation via global star registration algorithm and in order for this technique to succeed , yes the read noise of the sensor is the key factor and the read noise of the A7s sensor is very low compared to other DSLRs.
What DLSR /sensor are you using?
I agree that the FPN is also an issue and this not a problem with this method because with this kind of dobson mount almost all the frames are not aligned… (I can get more than 200 pixels shift relative to the reference image during a session even with precise mount alignment).
Feel free to ask if you have more questions.
Thanks and regards,
Fabrice
Beautiful Image…
out of curiosity tried to do some selective denoising on it…dont think much critical detail was lost…but then I don’t know this genre of photography. Perhaps de-noising is a faux pas !
PS… if its rude or rule breaking to post this please let me know and ill remove it…Obviously this is the OP’s image and just wanted to see how much detail could be retained in relation to noise removed.
I won’t speak for the poster, but I tend to think it is bad form to edit someone’s image when they haven’t asked for it; that is my personal feeling.
Thanks for having improved the image with this technique, can you please provide more details about the tool/settings that you have used to achieve this result?
I don’t see any problem with posting new versions of the image as long as the original reference is keep
Many thanks,
Fabrice
No problem…the technique is essentially two steps.
- I denoise the image. I used Dfine in the NIK collection which was previously free while google was developing it (or rather not developing ). Have never seen denoising as good as this. I used the manual mode to adjust the result, push it up a bit.
- Use some form of masking to determine where the denoised bit should be used and where the original image should show through.
The masking here is to avoid the loss if detail in focal points, such as the galaxy at the centre which has no masking at all. Lens flares around the starts, which would turn to mush, and smaller starts which might disappear altogether given they are essentially indistinguishable from noise.
I used a few forms of masking here…
- luminosity mask so that portions of the image which are essentially black reveal the full denoised effect/image. This leaves the brighter portions of the image mostly unmasked and so more detailed. Critical setting here is the mask feathering which must be tweaked to blends smoothly (or not).
- manually painted mask in the centre focal point, which i feathered out along the cosmic dust halo with a soft edged brush.
The settings parameters are not really informative as they would be different for every image…and also need fine adjustment at 100% scale to see exactly whats happening…no secrets here. Very simple, just context dependent.
I spent 5 min max…a less rushed job could produce better results no doubt.
1 Like
Thanks @PhilipB for your detailed explanation.
used the same kind of noise reduction tools with Pixinsight with mask too with bright areas but the tradeoff between noise reduction, keeping details and not having a artificial result was difficult to achieve so I used the noise reduction very cautiously.
It seems that you managed to overcome this issue with the tool/settings that you have used, many thanks!
Fabrice
hello @fab4space just for the sake of curiosity: did you made these 19000 shoots blocking the shutter opened? is it possible on the sony A7? Otherwise, 19000 shoots would impact the shutter-life of your camera…
I don’t think the A7 can do that. The shutter life is mentioned at 200,000 but there are some out there with over 500,000
Some models (my A7Rii i.e.) have also electronic shutter option.
But one more curious question to the shooter: did you take the images in dark sky area? I can presume yes, or?
1 Like
Considering a middle way of 400k shoots, this single awesome picture cost 5% of the total shutter-life of his camera. I mean, one of the goal in my astrophotography hobby is to shoot one stunning picture like this, because I do really think it is perfect, but, does it really pay you for this 5% of camera life?
Btw: I can’t find the information I need about the electronic shutter, does it impact the total shutter-life too?
Hello,
What do you mean by “do not pull in the same direction”? Signal-to-noise-ratio will always be improved by adding up images (if properly weighted). The more the better.
@fab4space
Thanks a lot for the series showing how image quality is improving with the number of stacks! May I use these in my astronomy courses to demonstrate noise treatment by stacking? Berry & Burnell (2005) in their book on astronomical image processing show a similar series in black& white. Yours is much nicer!
Proper flat fielding should get rid of that, if the pattern is constant with time (otherwise it is not an FPN).
Hermann-Josef
1 Like
I don’t think the electronic shutter count will effect the mechanical shutter. I don’t know if they are counted on the images, for Nikon it does.
I don’t really care, but I don’t take that many images. If you plan those numbers a few times I suggest to get a camera with electronic shutter.
Hello,
@aster94 Yes all the shots were done in electronic shutter mode with shutter always opened during shooting session otherwise it would have greatly impacted the mechanical shutter lifetime. I am using the silent shooting setting with the burst continuous mode. The mechanical shutter open once at the start of the shotting session then close once at the end.
@st.raw , no it was not in a dark sky area, it was in my backyard which is suburban sky: I can the milky way but there is light pollution. On the Bortle scale I would say this is between 4 and 5:
2 Likes
Thanks for this information, once I will buy a Sony a7 I will surely try something like you did 
2 Likes
@Jossie thanks for your message. Yes of course you can use them, feel free to show them in your course. I would be interested if the courses are available online to get a link to your site. If needed I can prepare other stack with different number of images (<500 for example).
Thanks,
Fabrice
2 Likes
@fab4space Thanks a lot for your permission. Unfortunately the course is not online. And due to Corona it is now also postponed.
A really noisy image, i.e. a stack of only a few images (or perhaps also just a single image!) would be very instructive.
Best wishes
Hermann-Josef
1 Like
Hello @Jossie, Ok I will provide stack of fewer images and also a single one using the same postprocessing to show the stacking improvment.
Best wishes
Fabrice
Light pollution gets worse with longer exposures.
Shot noise gets better with longer exposures and longer integration time.
Read noise gets better with more frequent reads.
So there is an optimum exposure time where these meet and an optimum integration time beyond which the image doesn’t improve. For me 120-180 secs for about 4 hours is the best. I don’t improve with, say 10 hours.